Convergence for the Wang-Landau density of states

Abstract

The Wang-Landau method of estimating the density of states $g (E)$ $g (E)$ has become a powerful tool in statistical mechanics. Here it is shown that the distribution of random walkers sampled using an estimated density of states can always be used to improve the estimate. Specifically, this can be done without resorting to an auxiliary modification factor $f$ $f$ , which previously has been used to find $g (E)$ $g (E)$ self-consistently through a procedure that reduces $f$ $f$ incrementally toward unity. This straightforward approach is validated for multiple, independent random walkers.

Received 23 May 2011

DOI:https://doi.org/10.1103/PhysRevE.84.065702

Authors & Affiliations

G. Brown

Department of Physics, Florida State University, Tallahassee, Florida 32306, USA and Computational Science and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830, USA

Kh. Odbadrakh

Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830, USA

D. M. Nicholson

Computational Science and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830, USA

M. Eisenbach

National Center for Computational Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830, USA

Article Text

Click to Expand

References

Click to Expand

Physical Review E

covering statistical, nonlinear, biological, and soft matter physics